Open Subtitles Multilingual Translation
Train Sequential Neural Networks in Nine Languages
By Huggingface Hub [source]
About this dataset
This dataset provides an invaluable opportunity to train a neural network model to effectively and accurately translate text between an array of nine different languages, including Finnish, Hindi, Basque, Esperanto, French, Armenian, Bengali, Icelandic and Russian. Each language CSV file includes three columns: an ID column; a meta column which provides information about the source of the sentence; and finally a 'translation' column that contains the translated sentence. The aim is to build a dataset suitable for training models capable of mastering multilingual translation tasks in order to bridge gaps between languages. Train your model with this unique dataset today!
More Datasets
For more datasets, click here.
Featured Notebooks
- 🚨 Your notebook can be here! 🚨!
How to use the dataset
This dataset is a great resource for anyone looking to build a translation model using neural networks. Here is a guide on how to use it:
-
Download the appropriate .csv files for the languages you need from the Kaggle dataset.
-
The data comes in an easily accessible CSV file, with ID, meta and translation columns included in each row of data. The ID column consists of integer values that can be used to identify each row and create unique feature ignition labels when training your model, while the meta column contains information about where each sentence originated from, allowing you to quickly filter out any sentences with suspect origins if needed. The translation column should include both English translations as well as their foreign language equivalents per sentence (depending on which language you are working with).
-
To train your neural network model it's important that you have enough training data available and try different language-pairs related sub-set datasets if available before assembling your final full dataset for training later on down the road once all inputs are ready (if needed). This Kaggle set should provide sufficient sample sizes per individual language pair so proceed according appropriate after downloading whatever subsets needed from this main database here first.
-
Now it’s time to construct our input features vector sets for our neural network configuration/setup by gathering all relevant variables in separate lists/arrays depending on preferred coding method used later when setting up our NN architecture layer setups appropriately based off all gathered items (elements) contained inside their respective list(s)/array(s) generated previously by implementing these steps mentioned above accordingly prior first before doing anything requiring input variable providing relevant core information found initially inside this Primary Open Subtitle Database explored so far properly earlier until now prior to continuing ahead next further below progressively further soon onward next momentarily right straight away very shortly right afterwards verily literally afterwards manually immediately properly eventually orderly personally autonomously biologically etc fortuitously contemporaneously instantaneously automatically justly necessarily lastly rightly confidently quixotically thankfully digitally informatively thereby correspondingly conjecturally constructively alike remarkably consistently instinctually markedly freely liberally perhaps anecdotally feasibly undeniably dynamically promptly easily holistically fairly evidently continually spontaneously intrinsically adaptively pictorially expressively intuitively hopefully methodically rationally prophetically perspicuously naturally savagely progressively peculiarly responsively whimsically illustratively skilfully tenaciously swiftly mysteriously productively continuously electromagnetically agitatedly constantly accurately ingeniously busily purposefully eagerly curiously exuberantly aud
Research Ideas
- Creating a neural network to automatically translate texts from any of the 9 languages in this dataset into any other language.
- Developing an AI-powered chatbot that can reply in multiple languages that the users prefer.
- Building an automatic translation system with real-time video conversation capabilities for use by professionals such as interpreters and international translators
Acknowledgements
If you use this dataset in your research, please credit the original authors.
Data Source
License
License: CC0 1.0 Universal (CC0 1.0) - Public Domain Dedication
No Copyright - You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission. See Other Information.
Columns
File: en-hi_train.csv
| Column name |
Description |
| meta |
Contains information about the source of the sentence. (String) |
| translation |
Contains either a manual or machine generated translation of that specific sentence from its original language to another language. (String) |
File: bs-eo_train.csv
| Column name |
Description |
| meta |
Contains information about the source of the sentence. (String) |
| translation |
Contains either a manual or machine generated translation of that specific sentence from its original language to another language. (String) |
File: fr-hy_train.csv
| Column name |
Description |
| meta |
Contains information about the source of the sentence. (String) |
| translation |
Contains either a manual or machine generated translation of that specific sentence from its original language to another language. (String) |
Acknowledgements
If you use this dataset in your research, please credit the original authors.
If you use this dataset in your research, please credit Huggingface Hub.